This proposal describes the research projects that the applicant will conduct and the applicant's plans for professional development as a cancer prevention researcher during the proposed award period. The research will focus on etiologic studies of cancer in relation to exposure to chemicals, physical agents, and other occupational factors. The research will identify specific risk factors for which there is a strategy for prevention through the prevention of exposure or change in behavior. The applicant will broaden his background in analytic biostatistics and mechanisms of carcinogenesis, and will participate in the development of cancer prevention through the recognition and control of occupational exposures. Four studies are proposed: 1. A case-control study of 155 CNS tumor patients in the aircraft industry which will investigate the risks due to occupational exposures to specific chemicals, families of chemicals, and industrial processes. The factors under study include organic solvents, lubricants and machining fluids, polycyclic aromatic hydrocarbons, hydrazines, ionizing radiation, and work as machinist. 2. a case-control study of 350 CNS tumor patients among males 25-69 in the general population. The factors under study include specific chemicals, occupations, ionizing radiation and head trauma. This study will complement the first be addressing non-occupational factors as well as occupational factors. 3. A case-control study of 750 male and female colon cancer patients in relation to physical activity and diet. The emphasis is on the lifetime interaction between these two factors. This study developed from a more general approach to the use of variables derived from occupational data for hypothesis testing, which is described. 4. The use of tumor registry data for the systematic evaluation of cancer risk from occupational exposures. The applicant will develop an exposure classification system, systematically evaluate cancer risks for each exposure, and use empirical Bayes techniques to rank the associations to identify those that are more likely to be true.